The Rock Burst Hazard Evaluation Using Statistical Learning Approaches

The great threat and destructiveness brought by a rock burst make its prediction prevention crucial in engineering. hazard evaluation at project locations is an effective way of preventing since currently real-time not available. Since different control factors discrimination conditions were accepted conventional risk determination methods, the same area may produce conflicting results. In this study, Naive Bayes statistical learning models based on model prior distributions representing highly complicated nonlinear relationship between impact built to evaluate hazards. results suggested that Gaussian has strongest performance over four preset distributions. Combining mechanics parameters measured laboratory stress data collected sites, proposed was successfully employed kimberlite diamond mine Canada. exhibits robustness generalization evaluation, which can be generalized for similar engineering cases with enough supported data.